Hi,
I am a newbie in this group, and I am working on a matlab project and
the questions is as follows:
"Select a variety of images or audio signal must be long enough, e.g.,
at least 10,000 samples) V called the designing or training set of
signals. Use at least two different linear predictors to generate the
differential signals. Then, choose the Huffman method, the modified
Huffman method, or the arithmetic method to perform the entropy coding
for the differential signals. Also, test the designed entropy coding
scheme on a few new images signals that are not in the original
training set.
I am stuck and I have no clue how to start with it. could anybody give
me a hand on this?
Thank you very much.
Nick
Need help in Image processing matlab project
Started by ●December 7, 2005
Reply by ●December 8, 20052005-12-08
hi ,, this is bhupendra.......
linear prediction of the signal( may be 2D image or 1D audio) predicts the value of the signal.
suppose in the case of image..... the current value of the pixel can be estimated from the previous pixels values( may be top , left, left corner..) ...
so your linear prediction is a short of the estimate..... if your estimate is correct most of the time( means the prediction is good).... the diffrence between actual value & predicted one will be most of the time zero( or close to zero)....... so you can use huffmann coding ( pedefined table) to code this difference in less number of bits....... but the prediction is not good always.... so we use adaptive huffman coding.....or arithmetic coding..........
so read first about lossless coding of image... that uses a short of linear prediction...... read also huffman coding ... adaptive huffman coding... & arthemetic coding.......to get the whole idea of your problem.....////
hopes this will help cyrilkan <cyrilkan@cyri...> wrote: Hi,
I am a newbie in this group, and I am working on a matlab project and
the questions is as follows:
"Select a variety of images or audio signal must be long enough, e.g.,
at least 10,000 samples) V called the designing or training set of
signals. Use at least two different linear predictors to generate the
differential signals. Then, choose the Huffman method, the modified
Huffman method, or the arithmetic method to perform the entropy coding
for the differential signals. Also, test the designed entropy coding
scheme on a few new images signals that are not in the original
training set.
I am stuck and I have no clue how to start with it. could anybody give
me a hand on this?
Thank you very much.
Nick
linear prediction of the signal( may be 2D image or 1D audio) predicts the value of the signal.
suppose in the case of image..... the current value of the pixel can be estimated from the previous pixels values( may be top , left, left corner..) ...
so your linear prediction is a short of the estimate..... if your estimate is correct most of the time( means the prediction is good).... the diffrence between actual value & predicted one will be most of the time zero( or close to zero)....... so you can use huffmann coding ( pedefined table) to code this difference in less number of bits....... but the prediction is not good always.... so we use adaptive huffman coding.....or arithmetic coding..........
so read first about lossless coding of image... that uses a short of linear prediction...... read also huffman coding ... adaptive huffman coding... & arthemetic coding.......to get the whole idea of your problem.....////
hopes this will help cyrilkan <cyrilkan@cyri...> wrote: Hi,
I am a newbie in this group, and I am working on a matlab project and
the questions is as follows:
"Select a variety of images or audio signal must be long enough, e.g.,
at least 10,000 samples) V called the designing or training set of
signals. Use at least two different linear predictors to generate the
differential signals. Then, choose the Huffman method, the modified
Huffman method, or the arithmetic method to perform the entropy coding
for the differential signals. Also, test the designed entropy coding
scheme on a few new images signals that are not in the original
training set.
I am stuck and I have no clue how to start with it. could anybody give
me a hand on this?
Thank you very much.
Nick
